When utilizing the stepwise regression
method, all R2 were approximately 0.70 (range from 0.695 to 0.705).
When including all 18 color features, μG and σS were retained and R2
was 0.704. When only including the RGB color space, μG and σB were
retained and R2 was 0.703. When only including the HSI color space,
μI and σI were retained and R2 was 0.695. When only including the
L*a*b* color space, μL* and σa* were retained and R2 was 0.705.When
including RGB and HSI in combination, μG and σS were retained and
R2 was 0.704. When including RGB and L*a*b*in combination, μG and
σB were retained and R2 was 0.703.When including HSI and L*a*b* in
combination, μI and σa* were retained and R2 was 0.705. The results
of the linear and stepwise regression models showed the linear regression
method reached a higher R2 (0.83) compared to the stepwise regression
method (R2 = 0.70)..